bghira/SimpleTuner

Webhook spam: send_progress_update fires per-bucket during cache generation, triggers Discord 429s

#2645

· rafstahelin opened

on Mar 18, 2026

Qwen Image: collate_prompt_embeds crashes on variable-length captions in batched training

#2644

· rafstahelin opened

on Mar 18, 2026

I want to fine tune Lora-FLUX.2-klein-base-4B, but it's still downloading Mistral-3 text encoder from mistralai/Mistral-Small-3.1-24B-Instruct-2503...

#2643

· T-1009 opened

on Mar 17, 2026

Multiple simultaneous ST instances show error when shut down

#2642

· StableLlama opened

on Mar 14, 2026

[Feature Request] Self-Flow - Self-Supervised Flow Matching for Scalable Multi-Modal Synthesis (much better way to do REPA, works across all modalities, no external models required, outperforms every REPA method)

#2632

· MeiYi-dev opened

on Mar 5, 2026

[New model] Support LTX-2.3

#2631

· kabachuha opened

on Mar 5, 2026

validation using datasets error for flux kontext

#2627

· bghira opened

on Feb 21, 2026

[feature request] init from LoRA enhancement

#2626

· StableLlama opened

on Feb 20, 2026

Single safetensor file SDXL family training and validation not working

#2624

· Dunanwi opened

on Feb 17, 2026

Validation image generation stops during the training

#2623

· StableLlama opened

on Feb 14, 2026

Multi GPU Klein9B training early OOM

#2615

· StableLlama opened

on Feb 13, 2026

corrupted checkpoint is attempted to load

enhancement

good first issue

help wanted

#2613

· StableLlama opened

on Feb 13, 2026

Provide feedback

Saved searches

Use saved searches to filter your results more quickly

Webhook spam: send_progress_update fires per-bucket during cache generation, triggers Discord 429s

Qwen Image: collate_prompt_embeds crashes on variable-length captions in batched training

I want to fine tune Lora-FLUX.2-klein-base-4B, but it's still downloading Mistral-3 text encoder from mistralai/Mistral-Small-3.1-24B-Instruct-2503...

Multiple simultaneous ST instances show error when shut down

[Feature Request] Self-Flow - Self-Supervised Flow Matching for Scalable Multi-Modal Synthesis (much better way to do REPA, works across all modalities, no external models required, outperforms every REPA method)

[New model] Support LTX-2.3

validation using datasets error for flux kontext

[feature request] init from LoRA enhancement

Single safetensor file SDXL family training and validation not working

Validation image generation stops during the training

Multi GPU Klein9B training early OOM

corrupted checkpoint is attempted to load

Issues

Search results